
Three completely unrelated viruses independently figured out how to kill bacteria by attacking the same protein. Caltech scientists think this viral trick could unlock an entirely new class of antibiotics against drug-resistant superbugs.
Bacteria have been outsmarting our best antibiotics for decades. Turns out, viruses have been outsmarting bacteria for millions of years.
A team at the California Institute of Technology just figured out how they do it, and the discovery could crack open an entirely new class of antibiotics. Published in the February 26, 2026, issue of Nature, the study reveals that three completely unrelated viruses independently evolved to attack the exact same weak spot in bacteria. When evolution keeps arriving at the same answer, it's probably telling us something.
The viruses in question are bacteriophages (phages for short): tiny predators that infect and kill bacteria. They've been locked in an arms race with bacteria for billions of years, evolving tricks that make antibiotic chemists look like amateurs.
The Caltech team, led by graduate student Yancheng Evelyn Li and biochemistry professor Bil Clemons, used high-resolution imaging to study how phages kill their bacterial hosts. They focused on a class of viral weapons called single-gene lysis proteins, or Sgls. Think of these as protein-sized missiles, each coded by a single gene, designed to blow up a bacterium from the inside.
What they found was remarkable. Three different Sgls, from three phages with no evolutionary relationship to one another, all zeroed in on the same target: a bacterial protein called MurJ.
So what's MurJ, and why do viruses care about it so much?
Every bacterium wraps itself in a cell wall made of peptidoglycan, a mesh-like material that acts like chain mail armor. Without it, the bacterium can't hold its shape and dies. MurJ is the protein responsible for flipping the building blocks of that armor across the cell's inner membrane, like a conveyor belt feeding bricks to a construction crew. No flipping, no wall. No wall, no bacterium.
The viral Sgls work by jamming MurJ in one position, locking it like a stuck turnstile. High-resolution imaging showed the proteins wedge themselves into a groove on MurJ's surface, preventing the structural changes it needs to do its job. Construction stops. The bacterium falls apart.
Bristol Myers Squibb paid $800 million for a cancer drug from a Chinese biotech nobody was watching. It just hit both survival endpoints in the toughest form of breast cancer, marking the first time a bispecific ADC has pulled off that feat. The $8.4 billion bet is starting to look like a bargain.

Join thousands of biotech professionals who start their day with our free, daily briefing.
The kicker: MurJ exists only in bacteria. Human cells don't have it, don't need it, and wouldn't be affected by a drug targeting it. That's the holy grail of antibiotic design; a target you can hit without hurting the patient.
This is where the story gets really interesting. These three phages didn't copy each other's homework. They evolved completely independently, yet they all arrived at the same solution: disable MurJ.
Biologists call this convergent evolution. It's the same phenomenon that gave wings to both birds and bats despite their very different family trees. When nature stumbles onto something that works, it tends to stumble onto it repeatedly.
Clemons put it plainly: "This is the first strong evidence that evolution identifies MurJ as a great target for killing bacteria."
In other words, if you're looking for where to aim a new antibiotic, follow evolution's crosshairs.
The timing of this discovery is not lost on anyone working in infectious disease. The global antibiotic pipeline is in rough shape.
As of 2025, only 90 antibacterial agents sit in clinical development worldwide, down from 97 just two years earlier. Of those, a mere 11 qualify as truly innovative (meaning they work through new mechanisms that could overcome resistance). And only five target the WHO's "critical priority" pathogens, the deadliest drug-resistant bacteria on the planet.
Meanwhile, antimicrobial resistance already contributes to nearly 5 million deaths per year globally. Tens of thousands of those deaths happen in the U.S. alone. The pipeline is shrinking while the problem is growing; it's like watching a fire department sell off trucks during wildfire season.
No approved drug currently inhibits MurJ directly. That means the Caltech discovery doesn't just identify a new target. It identifies an untouched target, one that nature has already validated through billions of years of viral warfare.
Let's be clear about where we are on the journey from cool science to actual drugs: very early. The Caltech study is a fundamental discovery, not a clinical trial. Nobody's injecting phage proteins into patients yet.
But the path forward is more concrete than it might seem. Now that researchers know the exact groove on MurJ where these viral proteins bind, medicinal chemists can start designing small molecules that fit the same pocket. It's like having a locksmith study three different keys that all open the same door; once you understand the lock, you can make your own key.
The phage angle also hints at a deeper well of targets waiting to be found. Phage genomes are enormous libraries of anti-bacterial weapons, most of them unstudied. Screening those genomes for more Sgls could reveal additional vulnerabilities that bacteria can't easily evolve around.
This isn't happening in a vacuum. Phage-derived therapies have been gaining momentum after decades on the sidelines. The concept dates back to 1917, when French scientist Félix d'Hérelle first discovered bacteriophages. Phage therapy flourished in Eastern Europe and the Soviet Union but fizzled in the West once penicillin arrived.
Now, with resistance rendering many antibiotics useless, the West is taking phages seriously again. Phage cocktails (mixtures targeting bacteria through multiple mechanisms) are in early clinical testing. Phage-encoded enzymes called lysins, which punch holes in bacterial cell walls, represent another active area of development. The MurJ discovery adds a third lane to this revival: using phage-inspired insights to design entirely new chemical drugs.
That distinction matters. Phage therapy itself faces manufacturing and regulatory challenges because the "drug" is a living, evolving virus. But a small molecule designed to mimic what a phage protein does? That fits neatly into the existing drug development playbook.
The next steps will likely involve screening for more phage proteins that target MurJ (or similar essential bacterial proteins) and developing synthetic molecules that replicate their effects. The preclinical pipeline for antibacterials currently includes 232 programs, and discoveries like this one could feed promising new candidates into that funnel.
Will a MurJ inhibitor reach patients? It's years away at best. But the logic behind this approach is hard to argue with. Evolution spent billions of years testing targets. It keeps coming back to this one. Maybe it's time we listened.
Dupixent just landed its ninth FDA approval, this time as the first-ever drug for allergic fungal rhinosinusitis. The trial data is staggering: a 92% reduction in the need for surgery or steroids. For patients stuck in a brutal cycle of repeat sinus surgeries, everything just changed.